Mechanical and Microstructural Evaluation of Squeeze CastAl-4%Cu Alloy Using a Full-Factorial Experimental Design

KHAWAJA MUSTAFA AMIN HAIDER1,3 and NADEEM AHMAD MUFTI2,4

1.—Industrial and Mechanical Engineering Department, School of Engineering, University ofManagement and Technology (U.M.T), Lahore 54770, Pakistan. 2.—Department of Industrial andManufacturing Engineering, University of Engineering and Technology (U.E.T), Lahore 54890,Pakistan. 3.—e-mail: kh_mustapha@hotmail.com. 4.—e-mail: muftina@yahoo.com

A full factorial design was employed to investigate the effect of squeezepressure in conjunction with thermal parameters, i.e., melt and die temper-atures, on the mechanical properties of a squeeze cast Al-4%Cu alloy. Con-siderable variations in mechanical properties existed between different testruns, and these were discussed based on cooling rates previously quantified fora squeeze-cast Al-4%Cu alloy. The completeness of a full factorial design notonly identified a combination of process parameters for optimum results butalso facilitated an evaluation of the minimum pressure required to eliminateporosity and influence the die temperature on the microstructure of thesqueeze-cast alloy. In addition to the optimum run, particular importance wasgiven to those runs that had more desirable levels of control factors withrespect to energy consumption or tooling life. A microstructural analysis ofthese runs indicated the possibility of precipitation hardening that can openup further investigations toward the opportunities associated with in situ heattreatment of age-hardening, squeeze cast aluminum alloys.

INTRODUCTION

In the squeeze casting process, mechanical pres-sure is applied directly on melt by movement of punchin die cavity. This results in eliminating porosity,both shrinkage and gas related, as well as refininggrain structure due to higher cooling rates. Merits ofthis process have been substantiated by the amountof work reported in the literature to cover the effectsof different processing parameters, particularly ap-plied pressure, on the microstructure and mechanicalproperties of squeeze-cast alloys and metal-matrixcomposites (MMCs). However, increasing squeezepressure above a critical value, about 50 MPa, nee-ded to minimize porosity produced little improve-ment in the mechanical properties of wroughtaluminum alloys.1,2 Heat removal rates have beenreported to be high, but the effect on mechanicalproperties was not found to be significant over theeffect of an increase in density. The high rate of heatremoval can, however, result in mechanical quench-ing of an age-hardening aluminum alloy.3

Different process parameters, at different levelswithin the recommended range, have been investi-

gated generally using a single-factor approach toevaluate the resulting effect on the mechanicalproperties of squeeze-cast alloys and MMCs. Yang4

reported the effect of casting temperature onsqueeze cast aluminum and zinc alloys, while Su-kumaran et al.5 used an Al 2124 alloy and its MMCfor reporting the effect of squeeze pressure onmechanical properties. Yue2 reported a three-level,two-factor (32) experimental design involvingsqueeze pressure and casting temperature. Yongand Clegg6 reported a combined one-factor (squeezepressure) design with a 32 design (melt and dietemperatures) for a magnesium alloy.

This article attempts to investigate the effect ofpressure in conjunction with a change in thermalparameters, i.e., melt and die temperatures. Thenecessity of undertaking such research was basedon the premises that different thermal phenomenawere governing heat flow patterns in squeeze cast-ing at different levels of melt superheat.7 The cool-ing curves obtained in a previous research8

presented two very different profiles for squeezecasting for near-liquidus and high-superheat melttemperatures. An investigation into microstructure

JOM

DOI: 10.1007/s11837-014-0973-4� 2014 The Minerals, Metals & Materials Society

was also carried out to evaluate the processing-property relationship as well as any evidence ofin situ heat treatment of an Al-4%Cu alloy used inthis research.

Such considerations have not been appreciated inthe reported designs for evaluating the effect ofdifferent process parameters on mechanical prop-erties in squeeze casting. The one-factor-at-a-timeapproach used previously8 was therefore extendedusing a two-level, three-factor (23) full-factorialexperimental design for the characterization of asqueeze-cast Al-4%Cu alloy. This design will notonly give a combination of control factors withoptimum result but also provide informationregarding the feasibility of runs with more desirablelevels of control factors.

MATERIALS AND EXPERIMENTAL SETUP

The material used in this research was an alu-minum-copper alloy with a weight percentage of Al-Cu(3.817)-Mn(0.614)-Si(0.214)-Zn(0.163), which isclose to that of AA 2024 wrought composition. Thealloy was melted in an electric muffle furnace andskimmed to remove the oxide layer before it waspoured directly into the cast iron die block. The diecavity was cylindrical (2 in. diameter and 5 in.height) in shape as shown in Fig. 1, while pressur-ization of melt was carried out by a punch attachedto the ram of a 150-ton hydraulic press. Preheatingof the die-and-punch arrangement was achieved byan oxyacetylene torch.

The process parameters, or control factors,investigated in this research were applied pressure,pouring temperature, and die temperature as out-lined in Table I(a).

These factors and levels were selected because oftheir effects on cooling rate, energy requirements,and tooling life considerations. It is particularlyrelevant for the squeeze casting process because the

melt temperature for long freezing range (LFR) al-loys can be varied from a very high superheat,greater than 100�C, to a near-liquidus value. Thelevels for each of the two thermal parameters werethe same as previously investigated,8 while squeezepressure levels represented one value of 50 MPaconsidered sufficient for wrought aluminum alloysand a higher value of 130 MPa. The experimentaldesign for the eight runs is shown in Table I(b) andwas carried out with four replicates per runs,resulting in 32 total runs.

Fig. 1. Dimensions of the die used for squeeze casting.

Table I. 23 full-factorial experimental design forinvestigating the effect of process parameters onmechanical properties of squeeze cast Al-4%Cualloy

(a) Control factors with levels

Control factors

Levels of C.F.

Low (0) High (1)

Squeeze pressure (MPa) 50 130Melt temperature (�C) 650 825Die temperature (�C) Room temperature 200

(b) The design matrix

Samplerun

Squeezepressure

Melttemperature

Dietemperature

A 0 0 0B 1 0 0C 0 1 0D 1 1 0E 0 0 1F 1 0 1G 0 1 1H 1 1 1

Amin Haider and Mufti

Each squeeze-cast billet was sectioned to obtaintwo tensile test specimens as per ASTM B557M(Fig. 2). A Rockwell hardness test, scale F with 60-kgf load and 1/8-in. diameter ball, was performed onstubs that remained after machining the tensile testspecimen (Fig. 3). Microstructural analysis was alsocarried out for the run that showed maximummechanical properties as well as for those that werecomparable to the maximum run but with morefeasible levels of control factors.

RESULTS

The tensile test properties, i.e., ultimate tensilestrength (UTS) and percentage elongation, as wellas hardness test data have been summarized inTable II for the 23 full-factorial experimental de-sign.

Ultimate Tensile Strength (UTS)

UTS values for a squeeze-cast Al-4%Cu alloyshowed a considerable amount of variations amongthe eight test runs (Fig. 4), from a minimum of169 ± 13 MPa for run C (010) to a maximum of258 ± 24 MPa for run H (111). This maximum va-lue was higher than that reported for 2024 alloy,i.e., 220 MPa (max.), in the non-heat-treated con-dition but was not even comparable to those re-ported for heat-treated conditions being in the rangeof about 500 MPa.9 Runs G (011), A (000), and C(010) showed a minimum spread (range) of datavalues but with strength values much lower thanrun H. Run E (001) appeared more feasible with astrength value of 223 ± 10 MPa and conducted atlower values of melt temperature (650�C) and ap-plied pressure (50 MPa).

The optimum value of tensile strength observed inrun H is comparable to that reported by Hajjari andDivandari1 for 2024 alloy giving a tensile strength of250 MPa at an applied pressure of 70 MPa. Minget al.10 have however reported a higher strength

value of about 300 MPa at a pressure of 120 MPa,but the composition used—alloy 204.0—had a sig-nificantly higher amount of copper (�4.8%). Bothhave reported these values at approximately similarlevels of melt and die temperatures of about 750�Cand 250�C, respectively.

Percentage Elongation

The elongation values showed a somewhat similartrend of variations between the different sampleruns for the experimental design as was observed incase of tensile strength of a squeeze-cast Al-4%Cualloy (Fig. 5). Improvement in the elongation valuefor the optimum run indicated that a reduction inporosity levels and grain refinement must havecontributed to better tensile properties for thesqueeze-cast Al-4%Cu alloy. These values werehowever considerably lower than that generallyspecified for 2024 alloy (�12%) in the non-heat-treated condition.9 The sample used in tensile test-ing was based on a 4-mm-diameter sample while theelongation values for 2024 alloy have been specifiedfor a 20-mm-diameter sample. The tensile testsample size for certain processes, like casting, tendsto affect the resulting values of tensile properties,particularly elongation, ASTM B 557M-10 (sec.6.1.4 and 6.1.5), and the lower value for the opti-mized run can be attributed to this factor.

Run H (111) again showed the maximum value forelongation being 6% as compared to that of A (000)with a value of 2.9%. Run E (001) showed a some-what lower value of 4%, but it improved to nearlythe same as run H when pressure was increased,i.e., run F (101) with 5.9% elongation. A load-extension diagram for the eight sample runs areshown in Fig. 6, and it clearly indicates that thenature of fracture was highly brittle. There wasvery little evidence of any plastic deformation thatis generally associated with a 2024 alloy in the an-nealed condition. This can be attributed to the meltquality or inclusion of impurity in the form of oxidelayers that have found the way into die cavity be-cause of direct pouring. Hence, a minimum level of

Fig. 2. Tensile test specimen as per ASTM Standard B 557 M(dimensions in mm).

Fig. 3. Hardness testing specimen with highlighted areas indicatingvalues obtained from (a) near-surface, (b) subsurface, and (c) centralregions of the sample.

Mechanical and Microstructural Evaluation of Squeeze Cast Al-4%Cu AlloyUsing a Full-Factorial Experimental Design

elongation can be defined for ascertaining the levelof melt condition at pouring.11

Hardness

Hardness influences the machinability of an alloywith too-hard or too-soft materials considered adifficult prospect for machining due to tool or sur-face finish limitations. Pure aluminum is soft butaluminum-copper alloys have a harder matrix thatimproves their response to machining operations.Casting processes that are prone to porosity, how-ever, degrade this machining response as has beenobserved in the case of conventional high-pressuredie casting (HPDC).12 These HPDC processes passthe melt through a nozzle into the cavity resultingin a high velocity but equally turbulent melt entry.Excessive air entrapment results in a highly poroussubsurface structure of the cast product that is not

conducive for machining. Squeeze casting was ex-pected to eliminate porosity even for a thick-sec-tioned billet cast using a wrought composition, orLFR, alloy without any concern for loss of hard-ness.13

Yong and Clegg6 reported a scheme for hardnesstesting along cross section of a squeeze-cast sample,but the depth of sample was only 16 mm. A sampledepth of �110 mm allowed for studying variationsin hardness values due to a porous interior of a thicksection casting. Table II presents hardness databased on average values as received from threedifferent locations, i.e., close to (less than 5 mmbelow) the outer surface, subsurface (�15 mm), andclose to the center (core) of the sample. This isgraphically presented in Fig. 7, which indicateslower hardness values for locations away from theouter surface of the squeeze-cast samples. However,the variation between these hardness values for a

Table II. Tensile test properties and hardness data for the different test runs

Sample run

UTS Elongation Rockwell hardness (F scale) at±95% CI ±95% CI

MPa % Near surface ± 95% CI Middle surface Center

A 187 ± 11 2.9 ± 0.5 83.7 ± 3.7 80.6 79.6(000)B 190 ± 17 3.0 ± 1.1 85.7 ± 2.7 81.1 79.8(100)C 169 ± 13 3.3 ± 1.0 86.0 ± 2.5 83.8 82.0(010)D 201 ± 18 3.4 ± 1.0 87.8 ± 1.5 83.9 83.3(110)E 223 ± 13 4.0 ± 0.9 85.5 ± 3.6 80.6 81.4(001)F 228 ± 20 5.9 ± 1.4 87.2 ± 2.6 83.8 80.9(101)G 209 ± 10 3.1 ± 0.8 86.8 ± 2.0 82.6 81.6(011)H 258 ± 24 6.0 ± 1.6 88.0 ± 1.9 87.3 85.6(111)

CI = confidence interval.

Fig. 4. Graph of UTS values for the different test runs indicatingconsiderable variations in average values as well as data spread.

Fig. 5. Graph of percentage elongation values showing a somewhatsimilar trend as observed for UTS values.

Amin Haider and Mufti

given run did not indicate deterioration inmechanical properties with the removal ormachining of the outer hard skin, which is encoun-tered generally in metallic mold processes. Run H(111) was particularly exceptional in this regardshowing highly comparable values for the threedifferent locations selected for hardness testing.

Microstructure

The microstructure for the run H (111), whichshowed maximum values for the three mechanicalproperties investigated, mainly consisted of den-dritic growth patterns with a more or less cellularmorphology (Fig. 8a). Increasing the die tempera-ture must have lowered the cooling rate8 and gainedsome time for the solidifying dendrites to grow insize as can be observed in the micrographs. Thereare extended regions of more than 200 lm in size,along a given direction, with thick finger-like evi-dence of dendritic growth. At higher magnification

(Fig. 8b), there is an evidence of precipitation of thesecond phase in the primary or a-aluminum matrixfor the sample for run H. The tensile strength valuefor the run, however, was not found to be compa-rable to heat-treated values of Al-4%Cu alloys(�400 MPa) reported in the literature.9 This may bedue to the excessive precipitation, or over aging, ofthe alloy as clustering of second phase is reported tobe visible only at higher magnifications for the peakaged condition.14

In contrast to run H, the microstructure for run E(001) with the most desirable set of process param-eters consisted of a well-defined cellular patternthat had a predominantly globular morphology.However, there was not much evidence of precipi-tation of the second phase that must have resultedin comparatively lower properties than those ob-served for run H. It may also imply the retention ofa supersaturated structure of the primary matrixthat can be investigated for controlled precipitationat a more amenable temperature for aging. Thesecast structures for both runs E and H have beenobtained without applying water quenching at thetime of ejection of the squeeze-cast billet.

Fig. 6. Load-extension diagram for one set of samples for tensiletest.

Fig. 7. Graph showing differences in hardness values based onthree different locations for obtaining these values.

Fig. 8. Microstructure of sample H showing (a) dendritic growth pattern as well as evidence of (b) precipitation of the second phase.

Mechanical and Microstructural Evaluation of Squeeze Cast Al-4%Cu AlloyUsing a Full-Factorial Experimental Design

An increase in applied pressure, i.e., run F (101),made the properties comparable to run H except fortensile strength that showed little improvement.The grain refinement for both runs E and F waspractically of the same nature, i.e., globular as canbe observed in Fig. 9. This further substantiates thetheory that near-liquidus pressurization leads tograin refinement by undercooling of the melt withlittle influence of an increase in the heat-removalrates from the solidifying melt due to a better melt/mold contact.7

DISCUSSION

This section primarily focuses on the details pro-vided by the experimental design to evaluate infor-mation in the literature for different phenomenaassociated with the squeeze-casting process.

Mechanical Properties

Run C, which showed minimum strength, impliedthat increasing the pouring temperature tended todecrease strength at low values of squeeze pressureand die temperature. High melt temperatureaccompanied a greater liquid-to-liquid contractionof the melt in the die cavity, resulting in a consid-erable amount of loss in applied pressure before theonset of actual solidification. Higher pressure wasneeded to accommodate this increase in liquid-to-liquid contraction to achieve the expected increasein strength as observed for run D (110)(201 ± 18 MPa). The notion of critical value forsqueeze pressure therefore needs to be evaluatedagainst superheat requirements of the pouringoperation.

An increase in pressure from 50 MPa to 130 MPaat low values of thermal parameters, i.e., runs A(000) and B (100), produced a very small amount ofincrease in strength from 187 MPa to 190 MPa.This result was in line with the conclusions drawnby Hajjari and Divandari1 regarding an optimalvalue of applied pressure. However, at high values

of thermal parameters, i.e., runs G (011) and H(111), a significant increase from 209 ± 10 MPa to258 ± 24 MPa was observed. This increase wasgreater than that observed between runs C (010)and D (110) after compensating for melt contraction.Hence, for higher pressure level a combination ofthermal parameters can be investigated that pro-vided a more conducive heat removal pattern.

The undercooling effect has been presented inliterature as the main reason for improvement inmechanical properties of squeeze-cast alloys. Mostresearch7,13 has quoted the Clausius–Clapeyronequation to relate a change in freezing temperatureof a melt with an increase in applied load as givenbelow:

dp

dT¼ DS

DV¼ DH

TDV

This equation is, however, based on the conditionof equilibrium between the two states of liquid andsolid; i.e., the total change in free energy in chang-ing from one state to another is zero.15 The non-equilibrium cooling conditions associated with highcooling rates in contrast tend to lower the meltingpoint of the solidifying melt.3 A cooling curve ana-lysis of data obtained directly from the squeeze-casting process has clearly demonstrated this dis-crepancy while indicating a change in equilibriumsolidification range with high cooling rates.8 Themechanical properties are, therefore, expected tovary as these mechanical and thermal effects willinfluence the cooling rate in squeeze-cast Al-4%Cualloy as demonstrated by the values of themechanical properties obtained in this research.

Microstructure

Yue2 reported a somewhat similar increase ingrain size at a higher level of pressure for a squeeze-cast aluminum AA 7010 alloy as observed for run H(111) (see the section titled, ‘‘Microstructure’’ in the

Fig. 9. Micrographs of (a) sample E and (b) sample F showing a more or less equiaxed cellular structure with some evidence of second phaseprecipitation. For sample F, it is more profound in the form of globules that can also indicate overaging.

Amin Haider and Mufti

Results section) but with a resulting decrease inmechanical properties. This must be a result ofsolidification patterns in supercooled melt, i.e., arestriction of nucleation frequency in favor of den-dritic growth.15 It has been reported that highpressure in squeeze casting tends to increase thesolidification range, while a high die temperaturetends to impose a thermal arrest at the start ofsolidification.8 The cause of large grain size andlowering of properties must have been due to thiscombined effect of high pressure and die tempera-ture in spite of a high melt superheat (�150�C),which was intended to facilitate a high rate of heatremoval.

Maeng et al.16 attributed the improvement inmechanical properties not only to microstructuralrefinement but also to an increase in the solubilityof solute atoms as well as a change in morphologyand distribution of the intermetallic or eutecticphase. The effect of microstructural morphology cantherefore combine with that of a higher coppercontent in the primary matrix. This retention ofsupersaturated structure can later result in pre-cipitation of the second phase. The processing

scheme defined for run H therefore needs to befurther investigated for a possibility of in situsolution treatment for an Al-4% Cu alloy in partic-ular and other age-hardening compositions of alu-minum alloys in general. Solution treatment is ahigh-temperature processing step in the heattreatment of aluminum alloys that can considerablyincrease the cost of producing a high-strength alu-minum alloy product.3,9

Comparing sample runs A (000) versus E (001)and B (100) versus F (101), it can be observed fromTable II the increase in die temperature from a colddie to 200�C improved the mechanical properties ofthe squeeze-cast Al-4%Cu alloy. Such an increasewas also observed when the level of superheat forpouring was high, i.e., runs C (010) versus G (011)and D (110) versus H (111), again for both levels ofapplied pressure. These observations identifygreater prospects for die temperature to influencemechanical properties in squeeze casting mostlikely by affecting microstructural changes(Fig. 10). Vijian and Aruanachalam17 identified dietemperature as a significant control factor whileoptimizing the mechanical properties for LM24

Fig. 10. Micrographs showing shift in as-cast structure for (a) D versus H and (b) B versus F.

Mechanical and Microstructural Evaluation of Squeeze Cast Al-4%Cu AlloyUsing a Full-Factorial Experimental Design

aluminum alloy. The optimum die temperature re-ported was 150�C, which indicates that higher dietemperatures (300�C used by Vijian) have a ten-dency to increase grain size and decrease mechani-cal properties as have been postulated for the effectof an increase in pressure by Yue for his analysis ofsqueeze-cast AA7010 alloy.

CONCLUSION

(1) Using a full factorial experimental designyielded more precise information regarding theeffect of squeeze pressure in conjunction withthermal parameters on the mechanical proper-ties of a squeeze-cast Al-4%Cu alloy.

(2) An increase in pressure at near-liquidus value ofthe melt temperature did not produce anysignificant improvement in strength, but at ahigh value of super heat, the increase wasconsiderable.

(3) The die temperature has a greater influencethan that of a mere fine-tuning parameter onthe resulting tensile properties of a squeeze-castAl-4%Cu alloy both at near-liquidus and highsuperheat melt temperature values.

(4) Despite low melt preparation for the Al-4%Cualloy used, squeeze casting gave results that notonly showed good consistency but also werecomparable to a 2024 wrought alloy in anannealed temper.

(5) There exists a great evidence to explore thepossibility of in situ heat treatment with thesqueeze casting process because there is astrong evidence for second-phase precipitationin the primary-aluminum matrix as observed inthe microstructure for one of the samples.

ACKNOWLEDGEMENTS

The authors are grateful to the Higher EducationCommission (H.E.C), Pakistan, and University ofEngineering and Technology (U.E.T), Lahore(Pakistan), for making this research possible.

REFERENCES

1. E. Hajjari and M. Divandari, Mater. Des. 29, 1685 (2008).2. T.M. Yue, J. Mater. Process. Technol. 66, 179 (1997).3. W.R. Heine, C.R. Loper, and P.C. Rosenthal, Principles of

Metal Casting, 2nd ed. (New Delhi, India: Tata McGraw-HillInc., 1976).

4. L.J. Yang, J. Mater. Process. Tech. 140, 391 (2003).5. K. Sukumaran, K.K. Ravikumar, S.G.K. Pillai, T.P.D. Ra-

jan, M. Ravi, R.M. Pillai, and B.C. Pai, Mater. Sci. Eng. A490, 235 (2008).

6. M.S. Yong and A.J. Clegg, J. Mater. Process. Technol. 145,134 (2004).

7. M.R. Ghomashchi and A. Vikhrov, J. Mater. Process. Tech-nol. 101, 1 (2000).

8. K.M. Amin and N.A. Mufti, J. Mater. Process. Technol. 212,1631 (2012).

9. F.W. Smith, Principles of Materials Science and Engineer-ing, 3rd edn. (New York: McGraw-Hill Inc., 1986), pp. 524–528, 537.

10. Z. Ming, Z. Wei-wen, Z. Hai-dong, Z. Da-tong, and L. Yuan-yuan, Trans. Non-ferrous Met. Soc. China 17, 496 (2007).

11. J.R. Davies, eds., Tensile Testing, 2nd ed. (Materials Park,OH: ASM International, 2004), p. 129.

12. B. Peter, Foundry Technology, 2nd ed. (Oxford, U.K.: But-terworth-Heinemann, 2001), pp. 603–604.

13. T.M. Yue and G.A. Chadwick, J. Mater. Process. Technol.58, 302 (1996).

14. A. Russell and K.L. Lee, Structure-Property Relations inNon-Ferrous Metals (New York: Wiley, 2005).

15. J.A. Dantzig and M. Rappaz, Solidification (Lausanne,Switzerland: EPFL Press, 2009), pp. 31–34.

16. D.Y. Maeng, J.H. Lee, C.W. Won, S.S. Cho, and B.S. Chun,J. Mater. Process. Technol. 105, 196 (2000).

17. P. Vijian and V.P. Aruanachalam, Int. J. Adv. Manuf.Technol. 33, 1122 (2007).

Amin Haider and Mufti